Fault diagnosis in an optimized rolling bearing using an intelligent approach.

Bearings are the significant component in machinery applications to run a system; hence, the damage of bearings can stop a running machine. The bearing faults often arise because of poor tribological behavior and wrong lubrication. Several fault prediction models already exist, but if the data were too large, the conventional techniques took a wide time range, gaining very low accuracy. The present research work has focused on designing an efficient bearing faults detection system and optimized tribological characters to address this issue. Moreover, the proposed mechanism is named as Hybrid ant lion and African buffalo-based Modular Neural Frame. The dataset taken for estimating the fault is Case Western Reserve University bearing center. In addition, to gain the finest result, mineral oil is added as a lubricant to reduce friction and to enhance the bearing life. Here, the fitness of Ant lion and African buffalo was upgraded in the classification layer of the modular neural frame that helps to optimize the parameters and has improved the fault prediction ratio. In addition, the proposed model is implemented in the MATLAB R2018b environment. Finally, the parameters were calculated and compared with other models in terms of accuracy, wear, friction, precision, recall, time, F-score and have gained the finest results in all scenarios. [ABSTRACT FROM AUTHOR]